Direct Metagenomic Detection of Viral Pathogens in Nasal and Fecal Specimens Using an Unbiased High-Throughput Sequencing Approach

Nakamura, Shota; Yang, Cheng-Song; Sakon, Naomi; Ueda, Mayo; Tougan, Takahiro; Yamashita, Akifumi; Goto, Naohisa; Takahashi, Kazuo; Yasunaga, Teruo; Ikuta, Kazuyoshi; Mizutani, Tetsuya; Okamoto, Yoshiko; Tagami, Michihira; Morita, Ryoji; Maeda, Norihiro; Kawai, Jun; Hayashizaki, Yoshihide; Nagai, Yoshiyuki; Horii, Toshihiro; Iida, Tetsuya; Nakaya, Takaaki

The GS FLX system produces several million bases in one 7.5 h run [20] . The PicoTiterPlate device, physical divided into 16 regions, was used in this study, with 8 samples being loaded into 2 regions each. A single run yielded 15,298-32,335 (average 24,738) reads. Data analysis was basically performed according to the protocol we previously reported [20] . However, since the template cDNA for the high-throughput sequencing was prepared with random RT-PCR, an extra step was added to remove tag sequences in order to use the sequence data for BLASTN search analyses. Figure 2 shows the fraction of organisms (from which the sequences in the database were derived) that showed the best hits for the query sequences (E-value,10 25 ). To identify viral species including subtype, higher matches (E-value,10 240 ) for the sequence reads were selected (Table 1 ).